koala

You need two drives for the OS, four for data. Hetzner boxes are cheap with 2 drives, cost multiplies if you add any other.

I use LDAP auth, but no SSO or external mounts. Actually, I tested external mounts, but they gave me bad vibes, although they are interesting.

The other thing, I just run a preview generator application, no other plugins.

I was looking at the Proxmox graphs. Now, looking at iostat, r/s measured over 10s hovers between 0 and 0.20, with no visible effect of spamming reload on a Nextcloud URL. If you want me to run any other measurement command, happy to.

I like to live on the edge of time and therefore have the feeling that debian based distros (although being very stable) are too “old” for my liking.

Nowadays, with Flatpaks, so many software providing binaries, etc. this does not matter so much. If you want, you can even use something like Distrobox to have containers for tools using whatever bleeding edge distro you want, but still have a solid stable underpinning.

Debian also has more stuff than you would expect in backports. The main sticking point is yes, you'll be stuck in Debian 12's KDE until 13 comes out. But that might be sufficient for you?

(You could also use Debian Testing, which is basically a rolling release. But I'd consider stable first.)

I see some CPU and memory usage on my setup... but I don't even see any IO!

Literally, the IO chart for "week (maximum)" on Proxmox for my Nextcloud LXC container is 0, except for two bursts, of 3 hours of less each. (Maybe package updates?)

The PostgreSQL LXC container has some more activity (but not much), but that's backing Nextcloud and four other applications (one being Miniflux, which has much more data churn).

Huh, what?

I see in your link that that image has support for KasmVNC, which is great and you could use to make Emacs work...

But the whole point of VS Code is that it can run in a browser and not use a remote desktop solution- which is always going to be a worse experience than a locally-rendered UI.

I kinda expect someone to package Emacs with a JS terminal, or with a browser-friendly frontend, but I'm always very surprised that this does not exist. (It would be pretty cool to have a Git forge that can spawn an Emacs with my configuration on a browser to edit a repository.)

Eh, my Nextcloud LXC container idles at less than 4.5% CPU usage ("max over the week" from Proxmox). I use PostgreSQL as the backend on a separate LXC container that has some peaks of 9% CPU usage, but is normally at 5% too.

I only have two users, though. But both containers have barely IO activity.

Web-accessible Emacs? What are you using?

I keep everything documented, along with my infrastructure as code stuff. Briefly:

• Nextcloud
• Vaultwarden
• Miniflux
• My blog
• Takahe (a multi-domain) ActivityPub server
• My health tracker CRUD data entry
• https://alexpdp7.github.io/selfhostwatch/
• Grafana (for health stats and monitoring data from Nagios)
• Nagios
• FreeIPA/Ipsilon (SSO)

edit: plus a few things that do not have a web UI.

I was going to mention ZFS, but I suspect Raspberries are too weak for ZFS?

If you can use ZFS in both sides, send/receive is the bomb. (I use it for my backups.) However, I'm not sure how well encryption would work for your purpose. IIRC, last time I looked at it, if you wanted an encrypted replica, the source dataset should be encrypted, which did not make me happy.

I'd love to work on making NASes "great" for non-technical people. I feel it's key. Sending encrypted backups through peers is one of my personal obsessions. It should be possible for people to buy two NAS, then set up encrypted backups over the Internet with a simple procedure. I wish TrueNAS Scale enabled that- right now it's the closest thing that exists, I think.

The next TrueNAS Scale can do LXC containers using Incus. It's similar to a VM, but more lightweight. You can create a container for any Linux distro and install Borg on that. With previous versions, I googled and found some instructions to run Borg in a container with SSH, or you could use a VM.

Borg also supports dummy SSH targets, that TrueNAS can provide. Apparently, it's lower performance-

Why the choice of TrueNAS Scale? For just a Borg target, you could run any Linux distribution.

YunoHost is a non-profit. Things could change, of course, but I'd fear more that YunoHost dies than it tries to monetize.

TrueNAS is backed by a for-profit company that so far has a good track record and looks pretty sustainable. Plus, while YunoHost might be a bit more troublesome, TrueNAS Scale is pretty much based around "open" things- their app catalog is basically Helm charts, for example.

Docker Compose is quite portable too, but if you are re-using YAML compose definitions from the Internet, or non-official container images by third-parties, there's also risks involved- not everything is easy to migrate! I prefer a very hands-on approach to my personal infra (I package some RPMs!), so I think I wouldn't personally use YunoHost, but I feel somewhat comfortable recommending it to others.